fix(web): content: filter now respects regex mode (#947)

* fix(web): ContentExpr now respects isRegexEnabled flag in query parser

Previously, content: prefix filters always used substring matching regardless
of the regex toggle, so patterns like content:next-auth.*beta searched for the
literal string "next-auth.*beta" rather than treating it as a regex.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* changelog

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: brendan-kellam

CHANGELOG.md

@@ -9,6 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 - Fixed search query parser rejecting parenthesized regex alternation in filter values (e.g. `file:(test|spec)`, `-file:(test|spec)`). [#946](https://github.com/sourcebot-dev/sourcebot/pull/946)
+- Fixed `content:` filter ignoring the regex toggle. [#947](https://github.com/sourcebot-dev/sourcebot/pull/947)
 
 ## [4.12.0] - 2026-02-26
 

packages/web/src/features/search/parser.ts

@@ -272,7 +272,15 @@ const transformTreeToIR = async ({
                 };
 
             case ContentExpr:
-                return {
+                return isRegexEnabled ? {
+                    regexp: {
+                        regexp: value,
+                        case_sensitive: isCaseSensitivityEnabled,
+                        file_name: false,
+                        content: true
+                    },
+                    query: "regexp"
+                } : {
                     substring: {
                         pattern: value,
                         case_sensitive: isCaseSensitivityEnabled,